Use and interpretation of composite end points in orthopaedic trials.

Abstract

Randomized controlled trials in orthopaedics are often underpowered to detect important differences in outcomes. Composite end points (CEPs) hold promise as a strategy to address this issue by combining multiple end points into one summary measure, thus increasing the observed event rate. The use of CEPs by trialists, however, can be problematic when they include components that vary greatly in importance to patients and when differences in apparent effect between components are large. We present an overview of CEPs with a focus on appropriate design and interpretation of results.